TAU Performance System(R) http://tau.uoregon.edu
Copyright 1997-2024 Department of Computer and Information Science, University of Oregon Advanced Computing Laboratory, Los Alamos National Laboratory Research Center Juelich, ZAM Germany
/ NOTE: PLEASE REFER TO tools/src/contrib/LICENSE files for open *
Instructions on installing TAU can be found in the INSTALL file. JAVA users should see README.JAVA
Tuning and Analysis Utilities: TAU
TAU is a program and performance analysis tool framework being developed for the DOE, ECP, and ASC program at University of Oregon. TAU provides a suite of static and dynamic tools that provide graphical user interaction and interoperation to form an integrated analysis environment for parallel Fortran, Python, C, and C++ applications. In particular, a robust performance profiling facility availble in TAU has been applied extensively in the ACTS toolkit. Also, recent advancements in TAU's code analysis capabilities have allowed new static tools to be developed, such as an automatic instrumentation tool. These two features of the TAU framework are described below.
TAU Portable Profiling Package
The model that TAU uses to profile parallel, multi-threaded C++, C, Fortran, UPC, Python, Chapel and Java programs maintains performance data for each thread, context, and node in use by an application. The profiling instrumentation needed to implement the model captures data for functions, methods, basic blocks, and statement execution at these levels. The instrumentation is complicated, however, by advanced features in the C++ language, such as templates and namespaces. All C++ language features are supported in the TAU profiling instrumentation, which is available through an API at the library or application level. The API also provides selection of profiling groups for organizing and controlling instrumentation. ACTS software layers have been instrumented and support for thread profiling has been recently added.
From the profile data collected, TAU's profile analysis procedures can generate a wealth of performance information for the user. It can show the exclusive and inclusive time spent in each function with nanosecond resolution. For templated entities, it shows the breakup of time spent for each instantiation. Other data includes how many times each function was called, how many profiled functions did each function invoke, and what the mean inclusive time per call was. Time information can also be displayed relative to nodes, contexts, and threads. Instead of time, hardware performance data can be shown. Also, user-level profiling is possible.
TAU supports GPUs from NVIDIA (CUDA/CUPTI), AMD (ROCm/Rocprofiler), and Intel (OneAPI/Level Zero) where a binary may be launched with tau_exec with no modification to the binary. It also supports OpenMP Tools Interface v5.0 (OMPT), the Kokkos profiling interface, MPI Tools Interface (MPI_T) and pthread.
TAU's profile visualization tool, paraprof, provides graphical displays of all the performance analysis results, in aggregate and per node/context/thread form. The user can quickly identify sources of performance bottlenecks in the application using the graphical interface. In addition, TAU can generate event traces that can be displayed with the Vampir trace visualization tool.
TAU is being developed jointly by the University of Oregon, Los Alamos National Laboratory, and Research Centre Jülich, ZAM, Germany. It can be found at:
http://tau.uoregon.eduThe work on extending TAU for OpenSHMEM and Kokkos regions was supported by the United States Department of Defense (DoD) and used resources of the Computational Research and Development Programs and the Oak Ridge Leadership Computing Facility (OLCF) at Oak Ridge National Laboratory.
We would like to thank the following sponsors for their support: U.S. Department of Energy and DOE national laboratories [www.doe.gov]. U.S. Department of Defense (HPCMP PETTT) [www.defense.gov]. National Science Foundation [www.nsf.gov]. French Alternative Energies and Atomic Energy Commission (CEA) [www.cea.fr]. NASA [www.nasa.gov]. ParaTools, Inc. [www.paratools.com]